<table border="1" cellspacing="0" cellpadding="8">
    <tr>
        <th>Issue</th>
        <td>
            <a href=https://github.com/llvm/llvm-project/issues/141025>141025</a>
        </td>
    </tr>

    <tr>
        <th>Summary</th>
        <td>
            [mlir][linalg] Re-organise tests for vectorization
        </td>
    </tr>

    <tr>
      <th>Labels</th>
      <td>
            mlir
      </td>
    </tr>

    <tr>
      <th>Assignees</th>
      <td>
            banach-space
      </td>
    </tr>

    <tr>
      <th>Reporter</th>
      <td>
          banach-space
      </td>
    </tr>
</table>

<pre>
    The current Linalg vectorization tests are split somewhat arbitrarily across multiple files:

```bash
ls -1 mlir/test/Dialect/Linalg/vectoriz*
mlir/test/Dialect/Linalg/vectorization-pad-patterns.mlir
mlir/test/Dialect/Linalg/vectorization-scalable.mlir
mlir/test/Dialect/Linalg/vectorization-unsupported.mlir
mlir/test/Dialect/Linalg/vectorization-with-patterns.mlir
mlir/test/Dialect/Linalg/vectorization.mlir
mlir/test/Dialect/Linalg/vectorize-conv-masked-and-scalable.mlir
mlir/test/Dialect/Linalg/vectorize-convolution-flatten.mlir
mlir/test/Dialect/Linalg/vectorize-convolution.mlir
mlir/test/Dialect/Linalg/vectorize-tensor-extract-masked.mlir
mlir/test/Dialect/Linalg/vectorize-tensor-extract.mlir
```

**Proposed improvements:**
1. Standardise naming: e.g., unify `vectorize` vs. `vectorization`.
2. Introduce a dedicated subdirectory (e.g., "Linalg/Vectorization") to eliminate repeated prefixes (`vectorize`/`vectorization`).
3. Standardise the test split between:
   * Tests using `transform.structured.vectorize` (with masking).
   * Tests using `transform.structured.vectorize_children_and_apply_patterns` (no masking).

Re 3., this split is already followed in "vectorization.mlir" and "vectorization-with-patterns.mlir", respectively.

**Specific steps:**
* **Rename:** "vectorize-tensor-extract-masked.mlir"  --> _**"vectorization-tensor-extract.mlir"**_  (it follows the format from _**"vectorization.mlir"**_).
* **Rename:** "vectorize-tensor-extract.mlir"  --> **_"vectorization-tensor-extract-with-patterns.mlir"_** (it follows the format from **_"vectorization-with-patterns.mlir"_**).
* **Merge** "vectorization-scalable.mlir" with other tests.
* **Update** convolution tests to follow the same naming and structural convention.

**Additional clean-up**
* Split "vectorization.mlir" and "vectorization-with-patterns.mlir" into smaller tests. At a minimum, extract non-Linalg operations into separate test files.
* Rename "vectorization-pad-patterns.mlir" to more clearly reflect that it tests `transform.apply_patterns.linalg.pad_vectorization`.

**Structure after the changes**
Without committing to all the fine details just yet, here is a possible end structure for the tests:
```bash
mlir/test/Dialect/Linalg/vectorization/
├── conv.mlir
├── conv-flatten.mlir
├── conv-with-patterns.mlir
├── extract.mlir
├── extract-with-patterns.mlir
├── insert-slice.mlir
├── insert-slice-with-patterns.mlir
├── linalg-ops.mlir
├── linalg-ops-with-patterns.mlir
├── pad.mlir
├── pad-with-patterns.mlir
├── unsupported.mlir
```
In addition to consistent naming, this structure provides a clear separation between tests using:
* `transform.structured.vectorize` - finer-grained tests, typically involving masking.
* `transform.structured.vectorize_children_and_apply_patterns` - broader pattern-based transformations, typically without masking.

**Next steps**

I’ll be uploading patches implementing these changes incrementally, and I’ll use this ticket as a reference for the overall direction. If you have any high-level comments, suggestions, or objections, please raise them here. For smaller details, let’s handle them through the individual PRs.

Thanks!

PRs:
* https://github.com/llvm/llvm-project/pull/140877
</pre>
<img width="1" height="1" alt="" src="http://email.email.llvm.org/o/eJysWF1v2zoS_TX0y0CCTH8_-ME3uQYC7C6KtLv7aIzEscSWIgWScur99QtSsmo7vkmcXsBAHYk8833m1OicLDXRms3-YJznqLGoEtdgQYxzNnscYesrY9fnb0a5Ecf1t4qgaK0l7eEfUqMq4UCFN1b-D700Gjw57wAtgWuU9OBMTS8VekCbS2_RSnUELKxxDupWedkogr1U5Nhkw7L4mWfdJ0dXsWyjHCRjqJW0jG8DPuPbR4mKivCt84Lx7ckPxgPIh49Ht5MGRdKg92S1S-PduyFcgQpzRZ-83mrXNo2xnsQnEV6kr34ziLtvUVIYfUhqdD9IJKjF59PQQRnVxmD2KgTyWYd6lE_c9qSdsQn99BYL3wf22zgngKG1-07n4fPFmsY4EiDrxpoD1aR9nIb4lmWbcQpfPWqBVkhHoLGWumSTDVBapow_QKvl_ghsng322TyDg0vPn8UCs3mWsmzDU3jS3hrRFgQIgoQs0JMA1-ZC2njjCIwvTxYY50OQ_7kA5JzxFXgDpGQtNXoCSw1FtMbSXv4kF5CuvGN8e8M3xlfBvcllwL6iyCs9peTkX4h0xxcAwPgGvkXaaZ3UZYjZW9Rub2ydOm_bwreWRHqRHMaXYV4gFDhkszP8GbRdUUklLOkdarHDplHH3WkKe0vaXNlh2eaZYBJT6yvp-sikA1SWUBxhb5QyL6EpdEj-jSnlHFCLVy9v0UCo0QNYcg0VXh5IHdPzBvzaUCH3sgDnqbnovJCM7o9n0ljT8O7c7JtDwzlAkrDJn7DrUa_9vTUpweFweBfqsZS-T4eLrRAqgR721tR_BXqFcsr6Z6K5DmOAfDOM21XYDebeiumvLLwF-SrCf5It6XV0t1YV5xBHwfiKbLfAL7H-3Qj0J7Azeu2XvTd9KDESh_WJomJ_nkYGVbxKOtbnvP02QsjwMJxQhDppm4sG_Bpn428YApDaG3A1KjUEChsPCLXUsm7rMCV9AUEbnfQKxzRkI6zrEahBG4guklIUL6eEdY312p_X-oLzkLjaWIpRW3UES_uwUMAHvSR9n94LArqkl1RFB9MGxe4Wzf-a8BNvAe59iD3IuAp1SW5I9X-lr0zroTB1Lb0P9fMGUKmuQaUmEORRKgffW-fhSD7kqyJLkbigMc7JXBHQWdljaw8MfhJ5VwrvPpESVke2YX9ytpqy1UP_ZZkNX2KjDSv3nXPXSuPd8zdl1lu3rjXA-0fvtyG1I-sTp2RB956_31rXdolp7jx9v6UGxQeP3Y99Q3Gf67MnDdgzUxiEwmgnnQ__8ekF2LC7h1YP8k0KCsMQZ_rEFAGiVy39UEdt0Q9D4NmPSJYkDqFNSotSk-gHKnhxbGSBSh1BBm4-hNHt9Ub6UQPvqZgEcmtQkIX-aZJj0KwDakeQl-689JRy7syJkv5FP30vOAayzzZPoTzLjK1WSkFO0DbKoAgBNeiLilwQySpK5EhQFbmByUDqwsZXwXjwJKyGC8Q2qknpwMviB3nAUCpLe7Kki19UZQ5kA_F1UjgsG3jaw9G0UOGBAPURKllWiaIDqUiYUbLzB3BtWZIbUmEsmPx7hxEfNIrQEVjsdW0d-TOFrbHDYupZNhxX5Af3HVSoheqv-cqatqyiu1ILeZCiRQVfnl2f5W8V6h-O8XH355dn96vdKu97nbdlfFtKX7V5Wpia8a1Sh9M_SWPN946Im1YpxrfjabZcLEZiPRGryQpHtB4vpotVNp_MZqNqPZvPc1qMF_l8WYwJ5ys-m02ni4WYTGg8mfORXPOMz7IZ59lqwiezdJrN8-l0sec5jadiv2TTjGqUKg32U2PLkXSupfV4Os74bKQwJ-X6ny1OEm_2OLLr6G_elo5NMyXDYh8QvPQq_tQRL8we2eyPjpHY7BGeKTG2RB3LEScz9MDFuhm1Vq3vzlj028WcRdcPa_7_AAAA___wwsPV">